A safe bet: How Covid-19 is fuelling investor interest in the datacentre industry

Doubling down on datacentre capacity

As more people use public cloud services, the more datacentre capacity the hyperscalers need to rapidly acquire to accommodate demand. The London colocation market is one of the biggest beneficiaries of this trend.

So much so that market tracking data released by the datacentre consultancy arm of global real estate consultancy Knight Frank in October 2020 said the London market is on course to become one of four European colocation hubs with an excess of 1,000mW of live power by 2023.

As more people use public cloud services, the more datacentre capacity the hyperscalers need to accommodate demand. The London colocation market, one of the biggest beneficiaries of this trend, is on course to become one of four European colocation hubs with an excess of 1,000mW of live power by 2023

And none of this has gone unnoticed by the investor community, says Stephen Beard, partner and head of Europe, the Middle East and Africa (EMEA) and Asia-Pacific datacentres at Knight Frank.

“The datacentre industry has been around for about 20 years, and if you look at performance over the past 10 years, datacentres have been running at around 10-12% year-on-year growth,” he tells Computer Weekly.

“In some ways, Covid-19 has benefited the market in terms of take-up and the reliance on remote working – and therefore datacentres – so return this year has been astronomical.

“Everyone is chasing clouds, which I know sounds romantic, but what we’ve seen over the past 12 months is more migration to public cloud, as every CEO or CFO [chief financial officer] bar none has been told to reduce costs,” he says.

“Naturally, Microsoft, Amazon and Google take more and more [colocation] leases across the globe, and that’s great news for the investors because they’re getting a triple-A grade covenant on that investment, which is squeezing the investment yield and further supporting returns.”

Knight Frank’s own data also shows that the main colocation hubs in Europe (which include London, Frankfurt, Amsterdam and Dublin) and Asia are on course to potentially double in size over the next three years, such is the level of pent-up demand for datacentre capacity in these areas.

While a large proportion of this demand can be attributed to the cloud-focused digital transformation activities of enterprises across the globe, the steady creep of digitisation into other areas of our lives as consumers is also playing its part, says Beard. 

“When you look at what else is happening in the world to further support the idea of continued digitisation, there aren’t many appliances now that aren’t smart. Everything is fundamentally going through a datacentre – every click and tap and swipe we make online goes through a datacentre,” he says.

“Then you look at the automotive trade, [which is] making up a significant proportion of customer take-up across the market, with the likes of Daimler, BMW and Volkswagen having deployed significantly into datacentres recently as part of a push to make the vehicles they manufacture more autonomous.”

Investor interest on the rise

Given the disruptive impact Covid-19 has had on the share prices of firms in the hospitality, bricks and mortar retail, and aviation industries, for example, datacentres are looking like an increasingly sure bet for investors seeking to broaden their investment portfolios.

“A lot of the potential investors in datacentres [are] coming in from other parts of real estate, which may be having some challenges, such as commercial property, for example, so some of those guys are looking to diversify,” says Steve Wallage, managing director of datacentre-focused analyst firm Danseb Consulting.

“There is a lot of opportunity. As well as the traditional, private equity investors and telecoms and IT-orientated infrastructure funds, we are seeing broader, less niche infrastructure funds showing interest in the market too,” he tells Computer Weekly.

“We’ve also seen pension funds take an interest, which is a reasonably new development over the past 12-18 months. They are clearly looking for a safer return and see datacentres as potentially a nice long-term receipt.”

“A lot of the potential investors in datacentres [are] coming from other parts of real estate, which may be having some challenges, so some of them are looking to diversify”

It is worth noting, however, that while Covid-19 has undoubtedly accelerated colocation take-up rates, demand was already hitting record highs in the years leading up to the pandemic. For this reason, its status as a desirable alternative asset class for investors has been growing for some time now, says Wallage.

For proof of that, one only has to look at how colocation market leader Equinix’s stocks are performing compared to those belonging to the “big tech” FAANG group of companies (Facebook, Amazon, Apple, Netflix and Google), he adds.

“Equinix’s market capitalisation is about $67bn, and Digital Realty is around $44bn, so – based on the scale and size of them – [datacentres] are effectively becoming an asset class in their own right, and Equinix has pretty much outperformed all of those wonderful FAANG stocks,” says Wallage.

Another reason why the industry is proving attractive to investors is the lengthy, 15- to 20-year lease terms that colocation tenants tend to sign, which brings a degree of predictability to these investments.

“Once a hyperscaler is in place in a datacentre, for example, they don’t tend to move unless they’re very unhappy or there are particular external factors – like property tax changes – within the region that means it no longer makes business sense to stay put,” says Wallage.

“There is also the hassle involved with moving once they’ve made a large investment in that datacentre, so there would certainly be some costs that come to bear should they move.”

But for investors looking for faster returns, there are other factors that dictate which operators they consider to be worth investing in, says Wallage.

“One thing investors are very keen to know is what the exit strategy is going to be,” he says. “How is this [colocation company] going to actually monetise?”

In those instances, they will be looking for a statement of intent around any potential plans the company might have to float on the stock market (has it not done so already) or position itself as a potential acquisition target for another player in the market, such as Equinix, Digital Realty or CyrusOne, for example.

This year alone has already seen a bumper crop of mergers and acquisitions (M&As) take place in the colocation market, with figures published in April 2020 by Synergy Research Group confirming this year as a record-breaker in terms of deal values just four months into it.

However, a datacentre operator being in the market to offload their facilities does not necessarily mean investors are guaranteed a good return just because there are high-value M&A deals going on elsewhere in the sector, cautions Wallage.

“We’ve seen sales processes for some datacentres where there has been more than 100 bidders or expressions of interest, so ridiculous levels of interest, but people are still wary of spending money on what are perceived to be ‘less good assets’,” he says.

In this category might be small-scale operators with a “couple of datacentres” in one European country, which tend to be a less appealing proposition to potential buyers than operators with sites in a number of countries with good, strong occupancy rates, he says.

“It’s very difficult for [an operator] with a greenfield site to get interest – they’ve got to get an anchor tenant in because investors are very wary of ‘unproven’ datacentres. An unproven datacentre with an unproven datacentre team managing it is the worst of all worlds,” he says.

“Also, market differentiation is important. If you are just another datacentre trying to operate in Slough, what are you really bringing to the party? That is a potential red flag for investors.”

The fact that investors are discerning when it comes to investing in this space is good news for the industry overall, says Wallage.  

“It is a good sign for the industry longer term that investors aren’t putting money into everything, and that they are being reasonably discerning,” he adds.

“One thing we’ve seen, though, is people who have not got great assets who assume this is a great time to sell, because demand for capacity is so high that they would get a ridiculously high valuation, but that’s not proven to be the case because there’s still a fair degree of discrimination.”

A tough nut to crack

On the back of these trends, it is little wonder that the datacentre sector has found itself in the crosshairs of the investor community, but it is not always an easy market for newcomers to break into.

Some of this stems from the veil of secrecy the sector operates under, which makes it something of an “opaque market” for investors to move into, and there is also a degree of confusion about whether datacentres should be considered a real estate or technology investment, says Wallage.

“Datacentres have never quite fitted into either camp completely,” he says.

Knight Frank’s Beard backs this point and says the technical aspects of running a datacentre can lead to unplanned expenditure when things go wrong, which can catch investors unawares.

“There are some risks that [investors] really need to understand in terms of the costs of replacing mechanical and engineering equipment in the event of a power outage, for example, which is a substantial, serious [form of] leakage.”

There are numerous examples, involving players who shall remain nameless, where datacentre development and investment deals have gone disastrously wrong because investors have not done enough to educate themselves about the market before getting stuck in, he continues.

“They’ve been sold the story about ‘this is where the world’s going, so you need to invest in this sector’, but they’ve done so with misconceptions that this is a sector that doesn’t employ huge amounts of people, and that [datacentres] can be location agnostic, which isn’t true,” he continues.

“Much like every other asset class, location is king because proximity to power and fibre connections is important. These are some of the fundamentals that investors have ignored in the past – and have lost out as a result.”

Flooding the market

More investor interest is good news for datacentre operators, and it stands to reason that this should translate into access to more readily available and flexible capital, but it could bring its downsides too.

“As the sector moves away from being considered an alternative asset class, and falls into line with the more traditional asset classes (because there is more capital being spent on datacentres than any other asset class), that means the market is opening up to new pools of capital,” says Beard. “And there are investors out there who are creating new operational platforms that are targeting much lower return thresholds.”

Historically, he says, the sector has seen a 15-20% internal rate of return (IRR) on investments, but “those days are gone” now that new investors are entering the fray who are happy to accept an IRR on their outlays of 8-10% because that is more in keeping with how their other “investment vehicles” perform.

“[They] can outbid the datacentre operators which are very stringent in that 15-20% [IRR] because that’s the business model, which is a big threat to operators today.”

Another factor that potential investors need to be aware of is the question that continues to hang over the industry regarding how much longer the continued heightened demand of the hyperscale community for colocation capacity will go on for.

Presently, the bulk of the demand for colocation capacity in London and the other major European datacentre hubs is coming from the hyperscale community. However, there is mounting evidence to suggest that some of the public cloud giants are looking to start building more of their own facilities, which could have a knock-on impact on how much colocation capacity they need in the future.  

“I suspect the investment community doesn’t quite understand the risks around the cloud providers building their own facilities and at scale. It is likely we will see a slowdown in [colocation] take-up as the cloud giants look to bring things in-house”

“I suspect the investment community doesn’t quite understand the risks at the moment around cloud providers building their own facilities and at scale,” says Beard.

“We’ve seen [evidence of it] recently in London with Google’s acquisition of a 30-acre site, so it is likely we will see a slowdown in [colocation] take-up as the cloud giants look to bring things in-house,” he adds.

“Historically, the hyperscalers haven’t been able to [self-build] from a capacity, time-to-market, resource or experience point of view, but over the past two or three years, the big three [Amazon, Microsoft and Google] have been gearing up to do this.”

They have been embarking on hiring sprees for real estate experts who “know the market through and through”, says Beard, and planning permission advisors to help them source suitable sites.

Despite this, Danseb Consulting’s Wallage is more sanguine about how the self-build trend is likely to play out for investors in the long term, given the time pressure many hyperscalers are under to meet the soaring demand for cloud services in new and existing markets.

“Clearly, that’s very difficult for them to do, so there remains a lot of pressure on them to still use colocation,” says Wallage. “It is likely more of the hyperscalers will go down the self-build route over time, but any shortfall in demand [for colocation capacity] will be met by others,” he continues.  

“When we talk about the ‘hyperscale demand’ now, we talk about Facebook, Amazon, Google, Microsoft and Alibaba, but in three to four years that community could include new players such as Zoom or [TikTok owner] ByteDance or Oracle or IBM. There are plenty of other companies out there that could potentially fill the gap.”